Purpose: We studied vincristine disposition after 169 weekly i.v. bolus injections in 32 children with acute lymphoblastic leukemia, non-Hodgkin lymphoma, or Wilms' tumor. The aim of the study was to determine intrapatient and interpatient variability in vincristine disposition and demographic, clinical, and biochemical characteristics influencing this variability.
Methods: Vincristine plasma concentrations were measured by a high-performance liquid chromatography assay with electrochemical detection. A limited sampling strategy was used based on a bayesian parameter estimation algorithm that is part of the ADAPT II software package. A two-compartment, first-order model was fitted to the data, and pharmacokinetic parameters were calculated from the model using the ADAPT II software. For statistical analysis, analysis of variance (ANOVA), t test, simple and multiple regression analysis, and non-parametric or robust equivalents were used.
Results: Results showed a large intrapatient and interpatient variability in distribution half-life, elimination half-life, total body clearance, apparent volume of distribution at steady state, and area under the concentration-time curve. Intrapatient variability was significantly smaller than interpatient variability for all these parameters except distribution half-life. The diagnosis or treatment protocol turned out to be the most predictive characteristic; leukemia and non-Hodgkin lymphoma patients had a significantly higher total body clearance than Wilms' tumor patients.
Conclusions: We conclude that both intrapatient and interpatient variability in vincristine pharmacokinetics is large in pediatric cancer patients and that variability, although significantly influenced by diagnosis, largely remains unpredictable.